Great to see this post, and the 3x improvement is very significant. Probably by coincidence, I am about two weeks into doing the same experiment myself! I currently employ a single person who I pay to sit beside me and watch me work for 8-10 hours a day (and 6 hours on the weekend).
I’m collecting similar data, perhaps I should also do a write-up after a month? Here’s a few differences in our setups:
My assistant’s primary job is to track how much time I’m spending on everything, and produce a google sheet at the end of the day showing what I did for different chunks (e.g. “10:30: Processing inbox + breakfast” or “13:34: Reading new Eliezer-Scott dialogue”).
At the end of the day, we look through the sheet together, and I label all the chunks into three categories: “Unendorsed”, “Necessary Busy Work”, “Priority Time”, and the sheet sums up the amount of time in each category. It’s aggregated in my master sheet, here’s a screenshot of what that looks like.
Perhaps obvious technical note: Each column had conditional formatting to color green the better times in that column. More time spent on priority is better, more time on busy work or undendorsed is worse.
Only today did I try having my assistant view my screen the entire time, and after a few hours I gave up, because there was so much sensitive content that I kept asking her to look away. In general I find it works very well just from her regularly asking and time-tracking what I’m working on.
There’s other things I’m doing differently. At the same time as starting this experiment I also wanted to try using lots of timers to force myself to get tasks done quickly. So either me or my assistant is constantly setting 5, 10, 15, 60 min timers, with sometimes a few mins of overtime if a task isn’t complete. (For the record this is my favorite such timer.) This also gives her a natural time to ask “and what task are you working on during this timer?”.
She also writes lots more detailed notes about what I do, that I later read back. This includes lots of observations that aren’t things I typically bring to awareness. There’s also a number of amusing or odd sentences in there, here a few I just found now looking through:
He is furiously typing in a slack channel (14:18)
“I wrote such a terrible essay,” -Ben “Want me to look at it?”-Ray “No it’s too bad,” -Ben
Ben returns, cream cheese in hand (17:46)
“huhooooh” “Holy shit” “Got the [data that I really wanted to analyze]” “Whatever I was doing before, I’m doing this now” (16:15)
Unauthorized facebook (18:21)
Looks around for a second then says “sometimes saying the problem out loud is enough” (12:06) but he didn’t tell me what it was and is just writing it down.
“This sucks,” Ben sighs, as timer goes off. (11:55)
Kurt comes in and talks about Ben’s German twin (11:49) [note: I do not have a German twin]
“I think I might be happier if I ate some food” aka breakfast (13:17)
“Oh no I didn’t check my email” “Well, too late another time” (13:24)
1-minute timer “Buy infinite sharpies” (16:39)
I had been working alone on projects for a while, and my basic hypothesis was that having a second person with whom to keep a social context about what I was working on, to check in regularly with, and who would be expecting me to do the things I just said I’d do, would improve my focus and time-spent.
I don’t have hard data on how much it’s improved my focus (I wasn’t tracking my time before hiring an assistant), but I’m quite confident my time-spent-working has substantially risen. I no longer spend big chunks of time slacking off which I used to, and because of the use of timers it’s far less often that I get into unproductive rabbit-holes that are “technically work” — after a timer goes off letting me know I’ve spent a whole 15 minute timer “checking slack”, I am much more likely to move on than to spend another hour.
Great to see this post, and the 3x improvement is very significant. Probably by coincidence, I am about two weeks into doing the same experiment myself! I currently employ a single person who I pay to sit beside me and watch me work for 8-10 hours a day (and 6 hours on the weekend).
I’m collecting similar data, perhaps I should also do a write-up after a month? Here’s a few differences in our setups:
My assistant’s primary job is to track how much time I’m spending on everything, and produce a google sheet at the end of the day showing what I did for different chunks (e.g. “10:30: Processing inbox + breakfast” or “13:34: Reading new Eliezer-Scott dialogue”).
At the end of the day, we look through the sheet together, and I label all the chunks into three categories: “Unendorsed”, “Necessary Busy Work”, “Priority Time”, and the sheet sums up the amount of time in each category. It’s aggregated in my master sheet, here’s a screenshot of what that looks like.
Perhaps obvious technical note: Each column had conditional formatting to color green the better times in that column. More time spent on priority is better, more time on busy work or undendorsed is worse.
Only today did I try having my assistant view my screen the entire time, and after a few hours I gave up, because there was so much sensitive content that I kept asking her to look away. In general I find it works very well just from her regularly asking and time-tracking what I’m working on.
There’s other things I’m doing differently. At the same time as starting this experiment I also wanted to try using lots of timers to force myself to get tasks done quickly. So either me or my assistant is constantly setting 5, 10, 15, 60 min timers, with sometimes a few mins of overtime if a task isn’t complete. (For the record this is my favorite such timer.) This also gives her a natural time to ask “and what task are you working on during this timer?”.
She also writes lots more detailed notes about what I do, that I later read back. This includes lots of observations that aren’t things I typically bring to awareness. There’s also a number of amusing or odd sentences in there, here a few I just found now looking through:
He is furiously typing in a slack channel (14:18)
“I wrote such a terrible essay,” -Ben “Want me to look at it?”-Ray “No it’s too bad,” -Ben
Ben returns, cream cheese in hand (17:46)
“huhooooh” “Holy shit” “Got the [data that I really wanted to analyze]” “Whatever I was doing before, I’m doing this now” (16:15)
Unauthorized facebook (18:21)
Looks around for a second then says “sometimes saying the problem out loud is enough” (12:06) but he didn’t tell me what it was and is just writing it down.
“This sucks,” Ben sighs, as timer goes off. (11:55)
Kurt comes in and talks about Ben’s German twin (11:49) [note: I do not have a German twin]
“I think I might be happier if I ate some food” aka breakfast (13:17)
“Oh no I didn’t check my email” “Well, too late another time” (13:24)
1-minute timer “Buy infinite sharpies” (16:39)
I had been working alone on projects for a while, and my basic hypothesis was that having a second person with whom to keep a social context about what I was working on, to check in regularly with, and who would be expecting me to do the things I just said I’d do, would improve my focus and time-spent.
I don’t have hard data on how much it’s improved my focus (I wasn’t tracking my time before hiring an assistant), but I’m quite confident my time-spent-working has substantially risen. I no longer spend big chunks of time slacking off which I used to, and because of the use of timers it’s far less often that I get into unproductive rabbit-holes that are “technically work” — after a timer goes off letting me know I’ve spent a whole 15 minute timer “checking slack”, I am much more likely to move on than to spend another hour.
Please do a write up as well. I think this experiment is very interesting and I’d love to read another report.
This is awesome! I highly encourage you to write up your experience; I think this should be more normalized!
Does this mean… Habryka? (does not seem very twin-like but does seem German)
No. Apparently there was a German fellow at some events that looked like me.